{"version":"2.1.273","anchor":"new-publish-is-still-settling-retry-path-for-artifact-depl","canonical_anchor":"publish-requests-are-now-safely-re-sent-once-if-the-response","heading":"Publish requests are retried once instead of failing on a settling error","tier":"notice","area":"Artifacts","scope":null,"heads_up":null,"url":"https:\/\/changelogs.core-directive.com\/v\/2.1.273\/e\/new-publish-is-still-settling-retry-path-for-artifact-depl","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.273","markdown":"### Publish requests are retried once instead of failing on a settling error\n\nA publish request that gets interrupted or hits a 'still settling' error is now automatically retried once instead of failing\n\n**Unclear.** Whether this behavior is enabled for any accounts is unknown, since the related gate `tengu_artifact_ask` is unread.\n\n**What**\n\nWhen publishing an artifact or deploy and the response never arrives, or the server responds with a new `publish_settling` error (HTTP 409), Claude Code now retries the publish request automatically instead of treating it as a failure.\n\n- A new retry wrapper (internally called `rf`) resends the interrupted or settling publish request exactly once.\n\n- The retried request is marked with a resend flag and an `attemptId`, so the server can recognize it as a retry of a publish that may have already landed, rather than duplicating it.\n\n**Why**\n\nThis avoids spurious publish failures and duplicate publishes when a deploy is still settling server-side or when the original response is lost in transit, making artifact\/page publishing more reliable.\n\n- Area: Artifacts\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 1\/5"}